OCR Software Recommendation

[i386]

I'm looking to scan in several hundred pages of tabular data into a spreadsheet that I'll be importing into a database application. This comes from one of the few vendors that doesn't offer this information in electronic format . To avoid time consuming manual entry I thought I'd try the OCR route. I've got a mid level HP scanner to test this on but need OCR. There are several popular brands of OCR software. Any recommendations?

 

[mscher]

Re: OCR Software Recommendation

I'm looking to scan in several hundred pages of tabular data into a spreadsheet that I'll be importing into a database application. This comes from one of the few vendors that doesn't offer this information in electronic format . To avoid time consuming manual entry I thought I'd try the OCR route. I've got a mid level HP scanner to test this on but need OCR. There are several popular brands of OCR software. Any recommendations?

Back in the day, Omnipage was about the best, but there there appears to be many more, these days. Some have free trials, so you might want to look at those.

Hopefully, your scan originals are square, have solid even type, and no spots, etc. Otherwise OCR can be real fun - not.

Are you planning to run OCR on the whole ducument page, or just part of it for indexing into the database?

Good luck.

 

[ThumbPkr]

Re: OCR Software Recommendation

I got an email about 10 days ago that Smith Micro was selling the Nuance Omnipage 16 for $99.99 if that is any help.It is supposed to be a 50 dollar saving.I have OCR capability with some of my scanner software but have not used it very much.If you are interested I will forward the link to you.Ron G

 

[arboldt]

Re: OCR Software Recommendation

Most PC-based scanners come with OCR software. Even though my scanner is several years old, it still does the occaisional job for me and I've seen no reason to upgrade it. The Omnipage OCR software also works for the most part. When I try to OCR a newpaper or magazine article, I sometimes think it'd actually be easier / faster to just retype it. Now I don't know how it'd do in tabular or spreadsheet sources. Often those have tiny footnotes etc that can throw of character recognition.

A bigger question would be about maintaining columnar separation. If there's a blank in Column C, it might just move whatever is in Column D over. That is, omitting a cell entry that should be blank can be worse than failing to recognize.

But there's something else here. You indicated you intend to scan several hundred pages. This would be a huge job for a PC-based scanner to the point I'd say there's got to be a better way.

At work, we have high-volume scanners for insurance claim forms, and specialized OCR software that has to be adapted for each field. That's not home-based at all.

I'd really think hard about what you're trying to accomplish. Can it even be viewed online? WOuld cut-and-past from a web page be better, as onerous as that would be?

 

[i386]

Re: OCR Software Recommendation

They only publish this information in book form which is ridiculous. These are medical codes for which all but the smallest medical practices will have to hand key into their electronic medical records software.

Before I took this job, they hand keyed these types of things in every time. Once I got here I've been getting these contracts in excel or csv format and importing them after massaging the columns. So far, this is the only one that's not offered in any electronic format. It boggles the mind.

My plan is to do a test at my desk with a few sheets. If it works, I'll set up the OCR software on one of our Fujitsu scanners. Its document feeder works pretty well and also scans fast.


I found out that the HP scanner I was referring to did come with some OCR software (I forgot which), but the CD has been lost. You can download drivers from HP, but not 3rd party bundled software (and understandably so). So, I have ordered a replacement CD from HP. It will be here next week. I'm going to try their software and go from there. If it works, I'll purchase a copy of the latest version to use on the other scanners. If it fails miserably I'm not sure what I'll do. Knowing which brand does the best with tabular data would be a plus.

Thanks for the posts. I'll let you know what happens.

 

[ThumbPkr]

Re: OCR Software Recommendation

I have Microtek scanners and I think that Omnipage is the software that is bundled with them.I am not sure if they have it on their website or if it would work with any other brand of scanner but I have it on CDs.Not saying any more than that.
The software that I use for scanning is called "VueScan".
That might be something that you would want to look into and it has its own OCR software application.Ed Hamrick is the developer and he is very good about helping with any problems you might have and his software is a labor of love which he updates regularly.Here is a link to his site.Ron G
http://www.hamrick.com/